[Paper Review] Optimal rate list decoding via derivative codes

Venkatesan Guruswami, Carol Wang|arXiv (Cornell University)|Jun 20, 2011

Coding theory and cryptography6 references5 citations

TL;DR

This paper introduces derivative codes—Reed-Solomon-like codes that encode a polynomial and its first m−1 formal derivatives—to achieve optimal list decoding radius approaching 1−R, where R is the code rate. The decoding uses a linear-algebraic approach involving multivariate interpolation and structured linear systems, enabling polynomial-time list decoding with a list size of O(1/ε²) for error fraction 1−R−ε, offering advantages in decoding with side information compared to folded Reed-Solomon codes.

ABSTRACT

The classical family of $[n,k]_q$ Reed-Solomon codes over a field $\F_q$ consist of the evaluations of polynomials $f \in \F_q[X]$ of degree $< k$ at $n$ distinct field elements. In this work, we consider a closely related family of codes, called (order $m$) {\em derivative codes} and defined over fields of large characteristic, which consist of the evaluations of $f$ as well as its first $m-1$ formal derivatives at $n$ distinct field elements. For large enough $m$, we show that these codes can be list-decoded in polynomial time from an error fraction approaching $1-R$, where $R=k/(nm)$ is the rate of the code. This gives an alternate construction to folded Reed-Solomon codes for achieving the optimal trade-off between rate and list error-correction radius. Our decoding algorithm is linear-algebraic, and involves solving a linear system to interpolate a multivariate polynomial, and then solving another structured linear system to retrieve the list of candidate polynomials $f$. The algorithm for derivative codes offers some advantages compared to a similar one for folded Reed-Solomon codes in terms of efficient unique decoding in the presence of side information.

Motivation & Objective

• To construct a family of codes that achieve the optimal trade-off between rate R and list decoding radius, approaching the theoretical limit of 1−R.
• To provide a constructive, polynomial-time list decoding algorithm for codes that can correct up to a fraction 1−R−ε of errors.
• To improve upon folded Reed-Solomon codes by enabling more efficient unique decoding when side information is available.
• To reduce list size from exponential to polynomial in s (e.g., O(1/ε²)) while maintaining optimal error-correction radius.

Proposed method

• Define (order m) derivative codes as evaluations of a polynomial f and its first m−1 formal derivatives at n distinct field elements over a large-characteristic field.
• Use multivariate polynomial interpolation to construct a candidate interpolation polynomial that vanishes at the received points with prescribed multiplicity.
• Formulate a structured linear system to recover the coefficients of the candidate message polynomial f from the interpolated polynomial.
• Apply a linear-algebraic decoding framework that avoids extension fields, enabling quadratic-time decoding complexity.
• Leverage side information (e.g., values of f and its derivatives at a random point α) to uniquely determine f without computing the full list.
• Use subspace-evasive sets to reduce list size from O(q^s) to O(s/ε) by restricting message space to low-density subsets.

Experimental results

Research questions

• RQ1 Can derivative codes achieve the optimal list decoding radius of 1−R−ε for any ε>0, matching the theoretical limit of random codes?
• RQ2 Is there a polynomial-time decoding algorithm for derivative codes that avoids costly root-finding in extension fields, as in previous folded RS constructions?
• RQ3 Can side information (e.g., derivative values at a random point) be used to enable unique decoding with high probability, avoiding the need to compute and prune a large list?
• RQ4 What is the trade-off between rate loss and list size reduction when using subspace-evasive sets in the code construction?
• RQ5 Can the decoding complexity be kept quadratic while achieving near-optimal error correction?

Key findings

• Derivative codes achieve a list decoding radius approaching 1−R, matching the optimal trade-off between rate and error correction, for any ε>0.
• The proposed decoding algorithm runs in quadratic time and avoids extension fields, unlike earlier folded RS decoding methods.
• With side information such as f(α), f′(α), ..., f^(s−1)(α) for a random α, unique decoding is possible with probability at least 1−nm/(sq), approaching 1 as q increases.
• By restricting messages to a subspace-evasive set of size q^(1−ζ)k, the list size is reduced from O(q^s) to O(s/ε), achieving O(1/ε²) list size for ε>0.
• The algorithm uniquely recovers f by solving a linear system after shifting by α, avoiding the expensive list computation step.
• The construction is randomized (Monte Carlo), but with high probability satisfies the list-decoding guarantees, offering a practical alternative to non-constructive random coding.